Table of contents

The behaviour of steam and its decomposition products in the ring discharge has been examined.

Dry hydrogen is not dissociated. The production of atomic hydrogen is dependent upon the presence of steam which dissociates into hydroxyl and atomic hydrogen. A secondary source of atomic hydrogen is then afforded by the interaction of hydroxyl with molecular hydrogen.

The escape from the discharge of atomic hydrogen, a long-lived species, favours the dissociation of steam. Mercury vapour, on the other hand, inhibits the formation of atomic hydrogen and thus leads to a high equilibrium steam concentration.

Unlike dry hydrogen, dry oxygen is dissociated into atoms, but these have a short life as such and recombine in the discharge to form molecular oxygen and ozone.

The reaction mechanisms occurring in the discharge are discussed in the light of spectrographic results.

A summarizing account is given of possible mechanisms which have been proposed to account for the emission of intense radio waves from the sun. Whilst most authors agree that the radiation from the undisturbed sun is due to an electron temperature of a million degrees in the solar corona, the greater intensity emitted from sunspots has led some authors to propose non-thermal mechanisms, based on the coherent oscillation of a large number of electrons in the solar corona above the sunspot. The mechanisms proposed for the maintenance of coherent oscillations are examined in detail in this paper and it is concluded that whilst they give adequate explanations for the oscillation observed in discharge tubes, no satisfactory mechanism has yet been suggested for the maintenance of electron oscillations in the solar corona.

It is therefore concluded that the observations can only be accounted for by the occurrence, in the solar corona near sunspots, of electron temperatures of up to 10¹⁰ deg. K.

The problem of maintaining an electron temperature of 10⁶ deg. K. in the normal corona, and 10¹⁰ deg. K. in the corona near a large sunspot, has been considered in an earlier paper; it was there shown that by the action of the general magnetic field of the sun and the non-uniform rotation of the solar surface, a potential difference is developed between the poles and the equator which is capable of maintaining an adequate electron temperature in both regions.

Several authors have suggested the possibility that cosmic rays are due to the acceleration of charged particles in the atmospheres of certain stars. Recent observations at radio frequencies have provided some evidence in support of this hypothesis which is reviewed in this paper.

The observations have shown that at least part of the radio-frequency radiation from the galaxy is emitted by discrete sources of small angular diameter. Measurements of the intensity of the radiation from these sources, and of the time variations of the intensity, indicate that the sources emit radio waves as if they were at a temperature of more than 10¹⁴ deg. K. It is concluded that this intense radiation cannot be caused by the coherent oscillation of a large number of electrons, but must be due to a genuine electron temperature of about 10¹⁴ deg. K. (corresponding to a mean electron energy of 10¹⁰ electron volts). The existence of this electron temperature indicates that in certain stellar bodies there are mechanisms capable of accelerating particles to cosmic ray energies.

A previous theory of the emission of radio waves from the sun is extended to other stellar bodies. It is concluded that a star having a surface magnetic field strength and a peripheral velocity somewhat greater than those which have so far been observed could emit radio waves having the observed intensity. A star of this type could also accelerate charged particles to cosmic ray energies.

The small visual brightness of the sources of radio waves suggests that these bodies may be characterized by small visual opacity or small photospheric temperature.

It is suggested that the difference between the angular distribution of cosmic rays and of the radio waves from the galaxy is due to the deflection of cosmic ray particles by magnetic fields. In addition to previous theories of a general magnetic field in the galaxy, it is shown that the scattering produced by the magnetic fields of the distributed sources themselves should be sufficient to produce an isotropic distribution of cosmic rays at the earth even though the sources may show a marked concentration near the galactic equator.

The production of electron pairs on bombardment of Pb, Pt, Sn, Ag and Cu by γ-rays from 200 mg. of Ra and its equilibrium products has been investigated. 1,200 photographs were taken for each element and the yield was 100 pairs in Pb, 85 in Pt, 72 in Sn, 63 in Ag and 50 in Cu. The distribution of energy of the positrons in Pb and Pt follows the shape of the theoretical curve of Bethe and Heitler, but the position of the maximum of the energy distribution curves in Sn, Ag and Cu shows a discrepancy with that to be expected theoretically. The theoretical difference between the average energies of the positrons and electrons in the elements studied, when plotted against Z, is consistently higher than the experimental, though both can be approximately represented by straight lines. The angular distribution of the positrons and electrons with respect to the incident γ-rays agrees with the theory.

The interest in the diamagnetic properties of quinodimethane lies in the fact that this molecule can be considered as a prototype for quinone. The diamagnetic anisotropy, Δχ, of p-benzoquinodimethane has been calculated by the molecular orbital method following the treatment given by F. London in 1937. We find: Δχ_{p-benzoquinodimethane} ~ 0.3Δχ_benzene. Calculations of the diamagnetic anisotropy of quinone itself fail because the perturbation used here breaks down in this case, in which all the centres in the molecule are not identical.

The result obtained for the quinodimethane can be discussed in connection with the observations by K. Lonsdale and K. S. Krishnan on quinone. We conclude that although Δχ of quinone is much reduced compared with that for benzene, we need not postulate, to account for this observation, a complete or nearly complete localization of the π-electrons in this structure.

The possibilities of obtaining precise momentum wave functions for atoms by direct solution of the wave equation in momentum space are examined in some detail. An iterative method of approximating to the wave function is employed in computing the momentum distribution function for the helium atom. Although in this instance considerable accuracy is achieved, formidable difficulties arise in extending the calculations to more complicated atoms.

A new approach is made to the problem of computing accurate momentum distribution functions for electrons in molecules. The usual molecular orbital wave function (linear combination of atomic orbitals) has already been employed for this purpose but has proved inadequate in attempts to explain the width of the Compton profile. Using the method of a previous paper, the l.c.a.o. type of molecular orbital is formally corrected by iteration. In most cases it seems likely that the corrected orbital will be associated with a greater spread of the momentum distribution and will thus lead to a broader Compton profile: the broadening should become more marked with increasing delocalization of the electron and may apparently be very considerable. A complete calculation has confirmed these conclusions in the case of the hydrogen molecule ion: a preliminary treatment of the neutral molecule is given.

Corrections to 1949 Proc. Phys. Soc. A 62 269.